This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
regcomp.c: Relax some restrictions on optimizations for locale
[perl5.git] / handy.h
CommitLineData
a0d0e21e 1/* handy.h
a687059c 2 *
1129b882 3 * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1999, 2000,
da5d8dbb 4 * 2001, 2002, 2004, 2005, 2006, 2007, 2008, 2012 by Larry Wall and others
a687059c 5 *
6e21c824
LW
6 * You may distribute under the terms of either the GNU General Public
7 * License or the Artistic License, as specified in the README file.
8d063cd8 8 *
8d063cd8
LW
9 */
10
85e6fe83 11#if !defined(__STDC__)
378cc40b
LW
12#ifdef NULL
13#undef NULL
14#endif
a687059c
LW
15#ifndef I286
16# define NULL 0
17#else
18# define NULL 0L
19#endif
85e6fe83
LW
20#endif
21
24792b8d
NC
22#ifndef PERL_CORE
23# define Null(type) ((type)NULL)
954c1994
GS
24
25/*
ccfc67b7 26=head1 Handy Values
954c1994 27
3609ea0d 28=for apidoc AmU||Nullch
24792b8d 29Null character pointer. (No longer available when C<PERL_CORE> is defined.)
2307c6d0 30
954c1994 31=for apidoc AmU||Nullsv
24792b8d 32Null SV pointer. (No longer available when C<PERL_CORE> is defined.)
954c1994
GS
33
34=cut
35*/
36
24792b8d
NC
37# define Nullch Null(char*)
38# define Nullfp Null(PerlIO*)
39# define Nullsv Null(SV*)
40#endif
8d063cd8 41
641d3f0b
PP
42#ifdef TRUE
43#undef TRUE
44#endif
45#ifdef FALSE
46#undef FALSE
47#endif
48#define TRUE (1)
49#define FALSE (0)
50
cf3f0ffb
DM
51/* The MUTABLE_*() macros cast pointers to the types shown, in such a way
52 * (compiler permitting) that casting away const-ness will give a warning;
53 * e.g.:
54 *
55 * const SV *sv = ...;
56 * AV *av1 = (AV*)sv; <== BAD: the const has been silently cast away
57 * AV *av2 = MUTABLE_AV(sv); <== GOOD: it may warn
58 */
59
b1bc3f34 60#if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
f065df04 61# define MUTABLE_PTR(p) ({ void *_p = (p); _p; })
b1bc3f34
NC
62#else
63# define MUTABLE_PTR(p) ((void *) (p))
64#endif
65
a062e10d 66#define MUTABLE_AV(p) ((AV *)MUTABLE_PTR(p))
ea726b52 67#define MUTABLE_CV(p) ((CV *)MUTABLE_PTR(p))
159b6efe 68#define MUTABLE_GV(p) ((GV *)MUTABLE_PTR(p))
dbebbdb4 69#define MUTABLE_HV(p) ((HV *)MUTABLE_PTR(p))
a45c7426 70#define MUTABLE_IO(p) ((IO *)MUTABLE_PTR(p))
b1bc3f34 71#define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p))
27d4fb96 72
bd31be4b
NC
73#ifdef I_STDBOOL
74# include <stdbool.h>
75# ifndef HAS_BOOL
76# define HAS_BOOL 1
77# endif
78#endif
79
8e84507e 80/* bool is built-in for g++-2.6.3 and later, which might be used
c1d22f6b
GS
81 for extensions. <_G_config.h> defines _G_HAVE_BOOL, but we can't
82 be sure _G_config.h will be included before this file. _G_config.h
8e84507e 83 also defines _G_HAVE_BOOL for both gcc and g++, but only g++
c1d22f6b
GS
84 actually has bool. Hence, _G_HAVE_BOOL is pretty useless for us.
85 g++ can be identified by __GNUG__.
86 Andy Dougherty February 2000
5d94fbed 87*/
3609ea0d 88#ifdef __GNUG__ /* GNU g++ has bool built-in */
5d94fbed 89# ifndef HAS_BOOL
c1d22f6b 90# define HAS_BOOL 1
5d94fbed 91# endif
5d94fbed
AD
92#endif
93
641d3f0b
PP
94/* The NeXT dynamic loader headers will not build with the bool macro
95 So declare them now to clear confusion.
96*/
8f1f23e8 97#if defined(NeXT) || defined(__NeXT__)
641d3f0b
PP
98# undef FALSE
99# undef TRUE
100 typedef enum bool { FALSE = 0, TRUE = 1 } bool;
101# define ENUM_BOOL 1
102# ifndef HAS_BOOL
103# define HAS_BOOL 1
104# endif /* !HAS_BOOL */
8f1f23e8 105#endif /* NeXT || __NeXT__ */
641d3f0b 106
5d94fbed 107#ifndef HAS_BOOL
61bb5906 108# if defined(UTS) || defined(VMS)
5d94fbed
AD
109# define bool int
110# else
111# define bool char
112# endif
c1d22f6b 113# define HAS_BOOL 1
a687059c 114#endif
0d3e774c 115
f2338a2e
DM
116/* a simple (bool) cast may not do the right thing: if bool is defined
117 * as char for example, then the cast from int is implementation-defined
18f5643b 118 * (bool)!!(cbool) in a ternary triggers a bug in xlc on AIX
f2338a2e
DM
119 */
120
18f5643b 121#define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0)
f2338a2e 122
46c6c7e2 123/* Try to figure out __func__ or __FUNCTION__ equivalent, if any.
e352bcff
JH
124 * XXX Should really be a Configure probe, with HAS__FUNCTION__
125 * and FUNCTION__ as results.
126 * XXX Similarly, a Configure probe for __FILE__ and __LINE__ is needed. */
46c6c7e2
JH
127#if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) || (defined(__SUNPRO_C)) /* C99 or close enough. */
128# define FUNCTION__ __func__
129#else
130# if (defined(_MSC_VER) && _MSC_VER < 1300) || /* Pre-MSVC 7.0 has neither __func__ nor __FUNCTION and no good workarounds, either. */ \
be73ebb0 131 (defined(__DECC_VER)) /* Tru64 or VMS, and strict C89 being used, but not modern enough cc (in Tur64, -c99 not known, only -std1). */
46c6c7e2
JH
132# define FUNCTION__ ""
133# else
134# define FUNCTION__ __FUNCTION__ /* Common extension. */
135# endif
136#endif
137
27d4fb96
PP
138/* XXX A note on the perl source internal type system. The
139 original intent was that I32 be *exactly* 32 bits.
140
141 Currently, we only guarantee that I32 is *at least* 32 bits.
142 Specifically, if int is 64 bits, then so is I32. (This is the case
143 for the Cray.) This has the advantage of meshing nicely with
144 standard library calls (where we pass an I32 and the library is
145 expecting an int), but the disadvantage that an I32 is not 32 bits.
146 Andy Dougherty August 1996
24fef2a7 147
dc45a647
MB
148 There is no guarantee that there is *any* integral type with
149 exactly 32 bits. It is perfectly legal for a system to have
150 sizeof(short) == sizeof(int) == sizeof(long) == 8.
693762b4 151
dc45a647
MB
152 Similarly, there is no guarantee that I16 and U16 have exactly 16
153 bits.
693762b4 154
8e84507e
NIS
155 For dealing with issues that may arise from various 32/64-bit
156 systems, we will ask Configure to check out
8175356b 157
3609ea0d
JH
158 SHORTSIZE == sizeof(short)
159 INTSIZE == sizeof(int)
160 LONGSIZE == sizeof(long)
dc45a647 161 LONGLONGSIZE == sizeof(long long) (if HAS_LONG_LONG)
3609ea0d 162 PTRSIZE == sizeof(void *)
dc45a647
MB
163 DOUBLESIZE == sizeof(double)
164 LONG_DOUBLESIZE == sizeof(long double) (if HAS_LONG_DOUBLE).
8175356b 165
27d4fb96
PP
166*/
167
69512466
JH
168#ifdef I_INTTYPES /* e.g. Linux has int64_t without <inttypes.h> */
169# include <inttypes.h>
dd0eed91
JH
170# ifdef INT32_MIN_BROKEN
171# undef INT32_MIN
172# define INT32_MIN (-2147483647-1)
173# endif
174# ifdef INT64_MIN_BROKEN
175# undef INT64_MIN
176# define INT64_MIN (-9223372036854775807LL-1)
177# endif
69512466
JH
178#endif
179
8175356b
JH
180typedef I8TYPE I8;
181typedef U8TYPE U8;
182typedef I16TYPE I16;
183typedef U16TYPE U16;
184typedef I32TYPE I32;
185typedef U32TYPE U32;
6b8eaf93
JH
186#ifdef PERL_CORE
187# ifdef HAS_QUAD
8175356b
JH
188typedef I64TYPE I64;
189typedef U64TYPE U64;
6b8eaf93
JH
190# endif
191#endif /* PERL_CORE */
8175356b 192
69512466 193#if defined(HAS_QUAD) && defined(USE_64_BIT_INT)
80524f33
JL
194# if defined(HAS_LONG_LONG) && QUADKIND == QUAD_IS_LONG_LONG
195# define PeRl_INT64_C(c) CAT2(c,LL)
196# define PeRl_UINT64_C(c) CAT2(c,ULL)
197# else
198# if LONGSIZE == 8 && QUADKIND == QUAD_IS_LONG
199# define PeRl_INT64_C(c) CAT2(c,L)
200# define PeRl_UINT64_C(c) CAT2(c,UL)
69512466 201# else
80524f33
JL
202# if defined(_WIN64) && defined(_MSC_VER)
203# define PeRl_INT64_C(c) CAT2(c,I64)
204# define PeRl_UINT64_C(c) CAT2(c,UI64)
69512466 205# else
80524f33
JL
206# define PeRl_INT64_C(c) ((I64TYPE)(c))
207# define PeRl_UINT64_C(c) ((U64TYPE)(c))
69512466
JH
208# endif
209# endif
e8c95190 210# endif
80524f33
JL
211# ifndef UINT64_C
212# define UINT64_C(c) PeRl_UINT64_C(c)
213# endif
214# ifndef INT64_C
215# define INT64_C(c) PeRl_INT64_C(c)
216# endif
e8c95190 217#endif
dc750f2e 218
d8668976 219#if defined(UINT8_MAX) && defined(INT16_MAX) && defined(INT32_MAX)
5ff3f7a4 220
5ff3f7a4
GS
221/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
222 Please search CHAR_MAX in perl.h for further details. */
223#define U8_MAX UINT8_MAX
224#define U8_MIN UINT8_MIN
225
5ff3f7a4
GS
226#define I16_MAX INT16_MAX
227#define I16_MIN INT16_MIN
228#define U16_MAX UINT16_MAX
229#define U16_MIN UINT16_MIN
230
5ff3f7a4
GS
231#define I32_MAX INT32_MAX
232#define I32_MIN INT32_MIN
0e983133
GS
233#ifndef UINT32_MAX_BROKEN /* e.g. HP-UX with gcc messes this up */
234# define U32_MAX UINT32_MAX
235#else
236# define U32_MAX 4294967295U
237#endif
5ff3f7a4
GS
238#define U32_MIN UINT32_MIN
239
240#else
241
5c9fa16e
KA
242/* I8_MAX and I8_MIN constants are not defined, as I8 is an ambiguous type.
243 Please search CHAR_MAX in perl.h for further details. */
27d4fb96
PP
244#define U8_MAX PERL_UCHAR_MAX
245#define U8_MIN PERL_UCHAR_MIN
79072805 246
27d4fb96
PP
247#define I16_MAX PERL_SHORT_MAX
248#define I16_MIN PERL_SHORT_MIN
249#define U16_MAX PERL_USHORT_MAX
250#define U16_MIN PERL_USHORT_MIN
79072805 251
c4f23d77 252#if LONGSIZE > 4
27d4fb96
PP
253# define I32_MAX PERL_INT_MAX
254# define I32_MIN PERL_INT_MIN
255# define U32_MAX PERL_UINT_MAX
256# define U32_MIN PERL_UINT_MIN
79072805 257#else
27d4fb96
PP
258# define I32_MAX PERL_LONG_MAX
259# define I32_MIN PERL_LONG_MIN
260# define U32_MAX PERL_ULONG_MAX
261# define U32_MIN PERL_ULONG_MIN
79072805
LW
262#endif
263
5ff3f7a4
GS
264#endif
265
58a9a5d5 266/* log(2) is pretty close to 0.30103, just in case anyone is grepping for it */
fc36a67e
PP
267#define BIT_DIGITS(N) (((N)*146)/485 + 1) /* log2(10) =~ 146/485 */
268#define TYPE_DIGITS(T) BIT_DIGITS(sizeof(T) * 8)
269#define TYPE_CHARS(T) (TYPE_DIGITS(T) + 2) /* sign, NUL */
270
ff68c719 271#define Ctl(ch) ((ch) & 037)
8d063cd8 272
3fe05580
MHM
273/*
274=head1 SV-Body Allocation
275
276=for apidoc Ama|SV*|newSVpvs|const char* s
277Like C<newSVpvn>, but takes a literal string instead of a string/length pair.
278
84bafc02
NC
279=for apidoc Ama|SV*|newSVpvs_flags|const char* s|U32 flags
280Like C<newSVpvn_flags>, but takes a literal string instead of a string/length
281pair.
282
3fe05580
MHM
283=for apidoc Ama|SV*|newSVpvs_share|const char* s
284Like C<newSVpvn_share>, but takes a literal string instead of a string/length
285pair and omits the hash parameter.
286
9dcc53ea
Z
287=for apidoc Am|void|sv_catpvs_flags|SV* sv|const char* s|I32 flags
288Like C<sv_catpvn_flags>, but takes a literal string instead of a
289string/length pair.
290
291=for apidoc Am|void|sv_catpvs_nomg|SV* sv|const char* s
292Like C<sv_catpvn_nomg>, but takes a literal string instead of a
293string/length pair.
294
7b4e7469 295=for apidoc Am|void|sv_catpvs|SV* sv|const char* s
3fe05580
MHM
296Like C<sv_catpvn>, but takes a literal string instead of a string/length pair.
297
9dcc53ea
Z
298=for apidoc Am|void|sv_catpvs_mg|SV* sv|const char* s
299Like C<sv_catpvn_mg>, but takes a literal string instead of a
300string/length pair.
301
7b4e7469 302=for apidoc Am|void|sv_setpvs|SV* sv|const char* s
3fe05580
MHM
303Like C<sv_setpvn>, but takes a literal string instead of a string/length pair.
304
9dcc53ea
Z
305=for apidoc Am|void|sv_setpvs_mg|SV* sv|const char* s
306Like C<sv_setpvn_mg>, but takes a literal string instead of a
307string/length pair.
308
309=for apidoc Am|SV *|sv_setref_pvs|const char* s
310Like C<sv_setref_pvn>, but takes a literal string instead of a
311string/length pair.
312
3fe05580
MHM
313=head1 Memory Management
314
315=for apidoc Ama|char*|savepvs|const char* s
316Like C<savepvn>, but takes a literal string instead of a string/length pair.
317
9dcc53ea
Z
318=for apidoc Ama|char*|savesharedpvs|const char* s
319A version of C<savepvs()> which allocates the duplicate string in memory
320which is shared between threads.
321
3fe05580
MHM
322=head1 GV Functions
323
324=for apidoc Am|HV*|gv_stashpvs|const char* name|I32 create
325Like C<gv_stashpvn>, but takes a literal string instead of a string/length pair.
326
327=head1 Hash Manipulation Functions
328
329=for apidoc Am|SV**|hv_fetchs|HV* tb|const char* key|I32 lval
330Like C<hv_fetch>, but takes a literal string instead of a string/length pair.
331
332=for apidoc Am|SV**|hv_stores|HV* tb|const char* key|NULLOK SV* val
333Like C<hv_store>, but takes a literal string instead of a string/length pair
334and omits the hash parameter.
335
510966aa
Z
336=head1 Lexer interface
337
338=for apidoc Amx|void|lex_stuff_pvs|const char *pv|U32 flags
339
340Like L</lex_stuff_pvn>, but takes a literal string instead of a
341string/length pair.
342
3fe05580
MHM
343=cut
344*/
345
65177ae0 346/* concatenating with "" ensures that only literal strings are accepted as argument */
26bfd924 347#define STR_WITH_LEN(s) ("" s ""), (sizeof(s)-1)
2efa8cc7 348
ba3a79e7
GA
349/* note that STR_WITH_LEN() can't be used as argument to macros or functions that
350 * under some configurations might be macros, which means that it requires the full
351 * Perl_xxx(aTHX_ ...) form for any API calls where it's used.
352 */
353
354/* STR_WITH_LEN() shortcuts */
355#define newSVpvs(str) Perl_newSVpvn(aTHX_ STR_WITH_LEN(str))
84bafc02
NC
356#define newSVpvs_flags(str,flags) \
357 Perl_newSVpvn_flags(aTHX_ STR_WITH_LEN(str), flags)
ba3a79e7 358#define newSVpvs_share(str) Perl_newSVpvn_share(aTHX_ STR_WITH_LEN(str), 0)
9dcc53ea
Z
359#define sv_catpvs_flags(sv, str, flags) \
360 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), flags)
361#define sv_catpvs_nomg(sv, str) \
362 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), 0)
363#define sv_catpvs(sv, str) \
364 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC)
365#define sv_catpvs_mg(sv, str) \
366 Perl_sv_catpvn_flags(aTHX_ sv, STR_WITH_LEN(str), SV_GMAGIC|SV_SMAGIC)
3fe05580 367#define sv_setpvs(sv, str) Perl_sv_setpvn(aTHX_ sv, STR_WITH_LEN(str))
9dcc53ea
Z
368#define sv_setpvs_mg(sv, str) Perl_sv_setpvn_mg(aTHX_ sv, STR_WITH_LEN(str))
369#define sv_setref_pvs(rv, classname, str) \
370 Perl_sv_setref_pvn(aTHX_ rv, classname, STR_WITH_LEN(str))
ba3a79e7 371#define savepvs(str) Perl_savepvn(aTHX_ STR_WITH_LEN(str))
9dcc53ea
Z
372#define savesharedpvs(str) Perl_savesharedpvn(aTHX_ STR_WITH_LEN(str))
373#define gv_stashpvs(str, create) \
374 Perl_gv_stashpvn(aTHX_ STR_WITH_LEN(str), create)
375#define gv_fetchpvs(namebeg, add, sv_type) \
376 Perl_gv_fetchpvn_flags(aTHX_ STR_WITH_LEN(namebeg), add, sv_type)
377#define gv_fetchpvn(namebeg, len, add, sv_type) \
378 Perl_gv_fetchpvn_flags(aTHX_ namebeg, len, add, sv_type)
379#define sv_catxmlpvs(dsv, str, utf8) \
380 Perl_sv_catxmlpvn(aTHX_ dsv, STR_WITH_LEN(str), utf8)
a038e571
NC
381#define hv_fetchs(hv,key,lval) \
382 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
383 (lval) ? (HV_FETCH_JUST_SV | HV_FETCH_LVALUE) \
384 : HV_FETCH_JUST_SV, NULL, 0))
385
386#define hv_stores(hv,key,val) \
387 ((SV **)Perl_hv_common(aTHX_ (hv), NULL, STR_WITH_LEN(key), 0, \
388 (HV_FETCH_ISSTORE|HV_FETCH_JUST_SV), (val), 0))
ba3a79e7 389
510966aa
Z
390#define lex_stuff_pvs(pv,flags) Perl_lex_stuff_pvn(aTHX_ STR_WITH_LEN(pv), flags)
391
b96d8cd9
NC
392#define get_cvs(str, flags) \
393 Perl_get_cvn_flags(aTHX_ STR_WITH_LEN(str), (flags))
5c1737d1 394
954c1994 395/*
ccfc67b7
JH
396=head1 Miscellaneous Functions
397
954c1994
GS
398=for apidoc Am|bool|strNE|char* s1|char* s2
399Test two strings to see if they are different. Returns true or
400false.
401
402=for apidoc Am|bool|strEQ|char* s1|char* s2
403Test two strings to see if they are equal. Returns true or false.
404
405=for apidoc Am|bool|strLT|char* s1|char* s2
406Test two strings to see if the first, C<s1>, is less than the second,
407C<s2>. Returns true or false.
408
409=for apidoc Am|bool|strLE|char* s1|char* s2
410Test two strings to see if the first, C<s1>, is less than or equal to the
411second, C<s2>. Returns true or false.
412
413=for apidoc Am|bool|strGT|char* s1|char* s2
414Test two strings to see if the first, C<s1>, is greater than the second,
415C<s2>. Returns true or false.
416
417=for apidoc Am|bool|strGE|char* s1|char* s2
418Test two strings to see if the first, C<s1>, is greater than or equal to
419the second, C<s2>. Returns true or false.
420
421=for apidoc Am|bool|strnNE|char* s1|char* s2|STRLEN len
422Test two strings to see if they are different. The C<len> parameter
423indicates the number of bytes to compare. Returns true or false. (A
424wrapper for C<strncmp>).
425
426=for apidoc Am|bool|strnEQ|char* s1|char* s2|STRLEN len
427Test two strings to see if they are equal. The C<len> parameter indicates
428the number of bytes to compare. Returns true or false. (A wrapper for
429C<strncmp>).
430
431=cut
432*/
433
8d063cd8
LW
434#define strNE(s1,s2) (strcmp(s1,s2))
435#define strEQ(s1,s2) (!strcmp(s1,s2))
436#define strLT(s1,s2) (strcmp(s1,s2) < 0)
437#define strLE(s1,s2) (strcmp(s1,s2) <= 0)
438#define strGT(s1,s2) (strcmp(s1,s2) > 0)
439#define strGE(s1,s2) (strcmp(s1,s2) >= 0)
440#define strnNE(s1,s2,l) (strncmp(s1,s2,l))
441#define strnEQ(s1,s2,l) (!strncmp(s1,s2,l))
378cc40b 442
36477c24
PP
443#ifdef HAS_MEMCMP
444# define memNE(s1,s2,l) (memcmp(s1,s2,l))
445# define memEQ(s1,s2,l) (!memcmp(s1,s2,l))
446#else
447# define memNE(s1,s2,l) (bcmp(s1,s2,l))
448# define memEQ(s1,s2,l) (!bcmp(s1,s2,l))
449#endif
450
568a785a 451#define memEQs(s1, l, s2) \
8e6497b6 452 (sizeof(s2)-1 == l && memEQ(s1, ("" s2 ""), (sizeof(s2)-1)))
568a785a
NC
453#define memNEs(s1, l, s2) !memEQs(s1, l, s2)
454
bbce6d69
PP
455/*
456 * Character classes.
457 *
458 * Unfortunately, the introduction of locales means that we
459 * can't trust isupper(), etc. to tell the truth. And when
460 * it comes to /\w+/ with tainting enabled, we *must* be able
461 * to trust our character classes.
462 *
463 * Therefore, the default tests in the text of Perl will be
464 * independent of locale. Any code that wants to depend on
465 * the current locale will use the tests that begin with "lc".
466 */
467
2304df62
AD
468#ifdef HAS_SETLOCALE /* XXX Is there a better test for this? */
469# ifndef CTYPE256
470# define CTYPE256
471# endif
472#endif
473
954c1994 474/*
ccfc67b7
JH
475
476=head1 Character classes
8a58bdcf
KW
477There are three variants for all the functions in this section. The base ones
478operate using the character set of the platform Perl is running on. The ones
479with an C<_A> suffix operate on the ASCII character set, and the ones with an
480C<_L1> suffix operate on the full Latin1 character set. All are unaffected by
181ab5fd 481locale and by C<use bytes>.
8a58bdcf
KW
482
483For ASCII platforms, the base function with no suffix and the one with the
484C<_A> suffix are identical. The function with the C<_L1> suffix imposes the
485Latin-1 character set onto the platform. That is, the code points that are
486ASCII are unaffected, since ASCII is a subset of Latin-1. But the non-ASCII
487code points are treated as if they are Latin-1 characters. For example,
488C<isSPACE_L1()> will return true when called with the code point 0xA0, which is
489the Latin-1 NO-BREAK SPACE.
490
491For EBCDIC platforms, the base function with no suffix and the one with the
492C<_L1> suffix should be identical, since, as of this writing, the EBCDIC code
493pages that Perl knows about all are equivalent to Latin-1. The function that
494ends in an C<_A> suffix will not return true unless the specified character also
495has an ASCII equivalent.
ccfc67b7 496
954c1994 497=for apidoc Am|bool|isALPHA|char ch
2787a470 498Returns a boolean indicating whether the specified character is an
06dc89af 499alphabetic character in the platform's native character set.
8a58bdcf
KW
500See the L<top of this section|/Character classes> for an explanation of variants
501C<isALPHA_A> and C<isALPHA_L1>.
502
503=for apidoc Am|bool|isASCII|char ch
504Returns a boolean indicating whether the specified character is one of the 128
505characters in the ASCII character set. On non-ASCII platforms, it is if this
506character corresponds to an ASCII character. Variants C<isASCII_A()> and
507C<isASCII_L1()> are identical to C<isASCII()>.
954c1994 508
954c1994 509=for apidoc Am|bool|isDIGIT|char ch
2787a470 510Returns a boolean indicating whether the specified character is a
06dc89af 511digit in the platform's native character set.
8a58bdcf 512Variants C<isDIGIT_A> and C<isDIGIT_L1> are identical to C<isDIGIT>.
954c1994 513
0c82b6df 514=for apidoc Am|bool|isLOWER|char ch
2787a470 515Returns a boolean indicating whether the specified character is a
0c82b6df 516lowercase character in the platform's native character set.
8a58bdcf
KW
517See the L<top of this section|/Character classes> for an explanation of variants
518C<isLOWER_A> and C<isLOWER_L1>.
0c82b6df 519
c99e91e9 520=for apidoc Am|bool|isOCTAL|char ch
2787a470 521Returns a boolean indicating whether the specified character is an
06dc89af 522octal digit, [0-7] in the platform's native character set.
8a58bdcf 523Variants C<isOCTAL_A> and C<isOCTAL_L1> are identical to C<isOCTAL>.
c99e91e9 524
0c82b6df 525=for apidoc Am|bool|isSPACE|char ch
2787a470 526Returns a boolean indicating whether the specified character is a
8a58bdcf
KW
527whitespace character in the platform's native character set. This is the same
528as what C<\s> matches in a regular expression.
529See the L<top of this section|/Character classes> for an explanation of variants
530C<isSPACE_A> and C<isSPACE_L1>.
0c82b6df 531
954c1994 532=for apidoc Am|bool|isUPPER|char ch
2787a470 533Returns a boolean indicating whether the specified character is an
06dc89af 534uppercase character in the platform's native character set.
8a58bdcf
KW
535See the L<top of this section|/Character classes> for an explanation of variants
536C<isUPPER_A> and C<isUPPER_L1>.
954c1994 537
8eea39dd
KW
538=for apidoc Am|bool|isWORDCHAR|char ch
539Returns a boolean indicating whether the specified character is a
540character that is any of: alphabetic, numeric, or an underscore. This is the
541same as what C<\w> matches in a regular expression.
542C<isALNUM()> is a synonym provided for backward compatibility. Note that it
543does not have the standard C language meaning of alphanumeric, since it matches
544an underscore and the standard meaning does not.
8a58bdcf
KW
545See the L<top of this section|/Character classes> for an explanation of variants
546C<isWORDCHAR_A> and C<isWORDCHAR_L1>.
547
548=for apidoc Am|bool|isXDIGIT|char ch
549Returns a boolean indicating whether the specified character is a hexadecimal
550digit, [0-9A-Fa-f]. Variants C<isXDIGIT_A()> and C<isXDIGIT_L1()> are
551identical to C<isXDIGIT()>.
8eea39dd 552
e7c1e6c1
KW
553=head1 Character case changing
554
954c1994 555=for apidoc Am|char|toUPPER|char ch
06dc89af
KW
556Converts the specified character to uppercase in the platform's native
557character set, if possible; otherwise returns the input character itself.
954c1994
GS
558
559=for apidoc Am|char|toLOWER|char ch
06dc89af
KW
560Converts the specified character to lowercase in the platform's native
561character set, if possible; otherwise returns the input character itself.
954c1994
GS
562
563=cut
353c9b6f 564
8a58bdcf 565Note that these macros are repeated in Devel::PPPort, so should also be
62fa66b6
KW
566patched there. The file as of this writing is cpan/Devel-PPPort/parts/inc/misc
567
954c1994
GS
568*/
569
7c062697
KW
570/* Specify the widest unsigned type on the platform. Use U64TYPE because U64
571 * is known only in the perl core, and this macro can be called from outside
572 * that */
573#ifdef HAS_QUAD
574# define WIDEST_UTYPE U64TYPE
575#else
576# define WIDEST_UTYPE U32
577#endif
578
3912bc88
KW
579/* FITS_IN_8_BITS(c) returns true if c doesn't have a bit set other than in
580 * the lower 8. It is designed to be hopefully bomb-proof, making sure that no
581 * bits of information are lost even on a 64-bit machine, but to get the
582 * compiler to optimize it out if possible. This is because Configure makes
583 * sure that the machine has an 8-bit byte, so if c is stored in a byte, the
584 * sizeof() guarantees that this evaluates to a constant true at compile time.
220c71bf
KW
585 */
586#define FITS_IN_8_BITS(c) ((sizeof(c) == 1) || !(((WIDEST_UTYPE)(c)) & ~0xFF))
cf301eb7 587
41f43cc2 588#ifdef EBCDIC
ff7eba81 589# define isASCII(c) (FITS_IN_8_BITS(c) && (NATIVE_TO_UNI((U8) (c)) < 128))
41f43cc2
KW
590#else
591# define isASCII(c) ((WIDEST_UTYPE)(c) < 128)
592#endif
593
ec79cfbb 594#define isASCII_A(c) isASCII(c)
3509bb53 595#define isASCII_L1(c) isASCII(c)
6515cb76 596
c2da0b36
KW
597/* The lower 3 bits in both the ASCII and EBCDIC representations of '0' are 0,
598 * and the 8 possible permutations of those bits exactly comprise the 8 octal
599 * digits */
3d33b3e3 600#define isOCTAL_A(c) cBOOL(FITS_IN_8_BITS(c) && (0xF8 & (c)) == '0')
c2da0b36
KW
601
602
ec79cfbb 603/* ASCII range only */
9fb1bf9d 604#ifdef H_PERL /* If have access to perl.h, lookup in its table */
181ab5fd
KW
605/* Bits for PL_charclass[]. These use names used in l1_char_class_tab.h but
606 * their actual definitions are here. If that has a name not used here, it
607 * won't compile. */
265c1f46
KW
608# define _CC_WORDCHAR_A 0
609# define _CC_SPACE_A 1
610# define _CC_DIGIT_A 2
611# define _CC_ALNUMC_A 3
612# define _CC_ALPHA_A 4
613# define _CC_ASCII 5
614# define _CC_CNTRL_A 6
615# define _CC_GRAPH_A 7
616# define _CC_LOWER_A 8
617# define _CC_PRINT_A 9
618# define _CC_PUNCT_A 10
619# define _CC_UPPER_A 11
620# define _CC_XDIGIT_A 12
621# define _CC_PSXSPC_A 13
622# define _CC_BLANK_A 14
623
624# define _CC_WORDCHAR_L1 15
625# define _CC_SPACE_L1 16
626# define _CC_ALNUMC_L1 17
627# define _CC_ALPHA_L1 18
628# define _CC_CNTRL_L1 19
629# define _CC_GRAPH_L1 20
630# define _CC_LOWER_L1 21
631# define _CC_PRINT_L1 22
632# define _CC_PUNCT_L1 23
633# define _CC_UPPER_L1 24
634# define _CC_PSXSPC_L1 25
635# define _CC_BLANK_L1 26
636# define _CC_IDFIRST_A 27
637# define _CC_IDFIRST_L1 28
638# define _CC_CHARNAME_CONT 29
639# define _CC_NONLATIN1_FOLD 30
640# define _CC_QUOTEMETA 31
e48bdfbc 641/* Unused: None
98d360d4 642 * If more are needed, can give up some of the above. The first ones to go
c2da0b36
KW
643 * would be those that require just two tests to verify; either there are two
644 * code points, like BLANK_A, or it occupies a single range like DIGIT_A,
645 * UPPER_A, and LOWER_A. Also consider the ones that can be replaced with two
646 * tests and an additional mask, so
647 *
648 * #define isCNTRL_A cBOOL(FITS_IN_8_BITS(c) \
649 * && (( ! (~0x1F & NATIVE_TO_UNI(c)])) \
650 * || UNLIKELY(NATIVE_TO_UNI(c) == 0x7f)))
651 *
652 * This takes advantage of the contiguous block of these with the first one's
653 * representation having the lower order bits all zero;, except the DELETE must
654 * be tested specially. A similar pattern can be used for for isCNTRL_L1,
655 * isPRINT_A, and isPRINT_L1
96ac0975
NC
656 */
657
658# ifdef DOINIT
659EXTCONST U32 PL_charclass[] = {
660# include "l1_char_class_tab.h"
661};
662
663# else /* ! DOINIT */
664EXTCONST U32 PL_charclass[];
665# endif
666
265c1f46
KW
667 /* The 1U keeps Solaris from griping when shifting sets the uppermost bit */
668# define _generic_isCC(c, bit_shift) cBOOL(FITS_IN_8_BITS(c) && (PL_charclass[(U8) NATIVE_TO_UNI(c)] & (1U << (bit_shift))))
4eeeb416
KW
669
670# define isALNUMC_A(c) _generic_isCC(c, _CC_ALNUMC_A)
671# define isALPHA_A(c) _generic_isCC(c, _CC_ALPHA_A)
672# define isBLANK_A(c) _generic_isCC(c, _CC_BLANK_A)
673# define isCNTRL_A(c) _generic_isCC(c, _CC_CNTRL_A)
674# define isDIGIT_A(c) _generic_isCC(c, _CC_DIGIT_A)
675# define isGRAPH_A(c) _generic_isCC(c, _CC_GRAPH_A)
676# define isIDFIRST_A(c) _generic_isCC(c, _CC_IDFIRST_A)
677# define isLOWER_A(c) _generic_isCC(c, _CC_LOWER_A)
678# define isPRINT_A(c) _generic_isCC(c, _CC_PRINT_A)
679# define isPSXSPC_A(c) _generic_isCC(c, _CC_PSXSPC_A)
680# define isPUNCT_A(c) _generic_isCC(c, _CC_PUNCT_A)
681# define isSPACE_A(c) _generic_isCC(c, _CC_SPACE_A)
682# define isUPPER_A(c) _generic_isCC(c, _CC_UPPER_A)
683# define isWORDCHAR_A(c) _generic_isCC(c, _CC_WORDCHAR_A)
684# define isXDIGIT_A(c) _generic_isCC(c, _CC_XDIGIT_A)
fb9bfbf7
KW
685 /* Either participates in a fold with a character above 255, or is a
686 * multi-char fold */
265c1f46 687# define _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c) ((! cBOOL(FITS_IN_8_BITS(c))) || (PL_charclass[(U8) NATIVE_TO_UNI(c)] & (1U << _CC_NONLATIN1_FOLD)))
4eeeb416 688# define _isQUOTEMETA(c) _generic_isCC(c, _CC_QUOTEMETA)
9fb1bf9d 689#else /* No perl.h. */
9fb1bf9d
KW
690# ifdef EBCDIC
691# define isALNUMC_A(c) (isASCII(c) && isALNUMC(c))
692# define isALPHA_A(c) (isASCII(c) && isALPHA(c))
693# define isBLANK_A(c) (isASCII(c) && isBLANK(c))
694# define isCNTRL_A(c) (isASCII(c) && isCNTRL(c))
695# define isDIGIT_A(c) (isASCII(c) && isDIGIT(c))
696# define isGRAPH_A(c) (isASCII(c) && isGRAPH(c))
697# define isIDFIRST_A(c) (isASCII(c) && isIDFIRST(c))
698# define isLOWER_A(c) (isASCII(c) && isLOWER(c))
699# define isPRINT_A(c) (isASCII(c) && isPRINT(c))
700# define isPSXSPC_A(c) (isASCII(c) && isPSXSPC(c))
701# define isPUNCT_A(c) (isASCII(c) && isPUNCT(c))
702# define isSPACE_A(c) (isASCII(c) && isSPACE(c))
703# define isUPPER_A(c) (isASCII(c) && isUPPER(c))
704# define isWORDCHAR_A(c) (isASCII(c) && isWORDCHAR(c))
705# define isXDIGIT_A(c) (isASCII(c) && isXDIGIT(c))
706# else /* ASCII platform, no perl.h */
707# define isALNUMC_A(c) (isALPHA_A(c) || isDIGIT_A(c))
708# define isALPHA_A(c) (isUPPER_A(c) || isLOWER_A(c))
709# define isBLANK_A(c) ((c) == ' ' || (c) == '\t')
ff7eba81 710# define isCNTRL_A(c) (FITS_IN_8_BITS(c) && ((U8) (c) < ' ' || (c) == 127))
2181da29 711# define isDIGIT_A(c) ((c) <= '9' && (c) >= '0')
9fb1bf9d
KW
712# define isGRAPH_A(c) (isWORDCHAR_A(c) || isPUNCT_A(c))
713# define isIDFIRST_A(c) (isALPHA_A(c) || (c) == '_')
714# define isLOWER_A(c) ((c) >= 'a' && (c) <= 'z')
715# define isPRINT_A(c) (((c) >= 32 && (c) < 127))
716# define isPSXSPC_A(c) (isSPACE_A(c) || (c) == '\v')
717# define isPUNCT_A(c) (((c) >= 33 && (c) <= 47) || ((c) >= 58 && (c) <= 64) || ((c) >= 91 && (c) <= 96) || ((c) >= 123 && (c) <= 126))
718# define isSPACE_A(c) ((c) == ' ' || (c) == '\t' || (c) == '\n' || (c) =='\r' || (c) == '\f')
2181da29 719# define isUPPER_A(c) ((c) <= 'Z' && (c) >= 'A')
9fb1bf9d 720# define isWORDCHAR_A(c) (isALPHA_A(c) || isDIGIT_A(c) || (c) == '_')
2181da29 721# define isXDIGIT_A(c) (isDIGIT_A(c) || ((c) >= 'a' && (c) <= 'f') || ((c) <= 'F' && (c) >= 'A'))
9fb1bf9d
KW
722# endif
723#endif /* ASCII range definitions */
ec79cfbb
KW
724
725/* Latin1 definitions */
8a58bdcf 726#ifdef H_PERL
4eeeb416
KW
727# define isALNUMC_L1(c) _generic_isCC(c, _CC_ALNUMC_L1)
728# define isALPHA_L1(c) _generic_isCC(c, _CC_ALPHA_L1)
729# define isBLANK_L1(c) _generic_isCC(c, _CC_BLANK_L1)
8a58bdcf 730/* continuation character for legal NAME in \N{NAME} */
4eeeb416
KW
731# define isCHARNAME_CONT(c) _generic_isCC(c, _CC_CHARNAME_CONT)
732# define isCNTRL_L1(c) _generic_isCC(c, _CC_CNTRL_L1)
733# define isGRAPH_L1(c) _generic_isCC(c, _CC_GRAPH_L1)
734# define isIDFIRST_L1(c) _generic_isCC(c, _CC_IDFIRST_L1)
735# define isLOWER_L1(c) _generic_isCC(c, _CC_LOWER_L1)
736# define isPRINT_L1(c) _generic_isCC(c, _CC_PRINT_L1)
737# define isPSXSPC_L1(c) _generic_isCC(c, _CC_PSXSPC_L1)
738# define isPUNCT_L1(c) _generic_isCC(c, _CC_PUNCT_L1)
739# define isSPACE_L1(c) _generic_isCC(c, _CC_SPACE_L1)
740# define isUPPER_L1(c) _generic_isCC(c, _CC_UPPER_L1)
741# define isWORDCHAR_L1(c) _generic_isCC(c, _CC_WORDCHAR_L1)
8a58bdcf
KW
742#else /* No access to perl.h. Only a few provided here, just in case needed
743 * for backwards compatibility */
744 /* ALPHAU includes Unicode semantics for latin1 characters. It has an extra
745 * >= AA test to speed up ASCII-only tests at the expense of the others */
746# define isALPHA_L1(c) (isALPHA(c) || (NATIVE_TO_UNI((U8) c) >= 0xAA \
747 && ((NATIVE_TO_UNI((U8) c) >= 0xC0 \
748 && NATIVE_TO_UNI((U8) c) != 0xD7 && NATIVE_TO_UNI((U8) c) != 0xF7) \
749 || NATIVE_TO_UNI((U8) c) == 0xAA \
750 || NATIVE_TO_UNI((U8) c) == 0xB5 \
751 || NATIVE_TO_UNI((U8) c) == 0xBA)))
752# define isCHARNAME_CONT(c) (isALNUM_L1(c) || (c) == ' ' || (c) == '-' || (c) == '(' || (c) == ')' || (c) == ':' || NATIVE_TO_UNI((U8) c) == 0xA0)
753#endif
754
755/* Macros for backwards compatibility and for completeness when the ASCII and
756 * Latin1 values are identical */
ec79cfbb
KW
757#define isALNUM(c) isWORDCHAR(c)
758#define isALNUMU(c) isWORDCHAR_L1(c)
8a58bdcf
KW
759#define isALPHAU(c) isALPHA_L1(c)
760#define isDIGIT_L1(c) isDIGIT_A(c)
761#define isOCTAL(c) isOCTAL_A(c)
762#define isOCTAL_L1(c) isOCTAL_A(c)
763#define isXDIGIT_L1(c) isXDIGIT_A(c)
764
765/* Macros that differ between EBCDIC and ASCII. Where C89 defines a function,
766 * that is used in the EBCDIC form, because in EBCDIC we do not do locales:
767 * therefore can use native functions. For those where C89 doesn't define a
768 * function, use our function, assuming that the EBCDIC code page is isomorphic
769 * with Latin1, which the three currently recognized by Perl are. Some libc's
770 * have an isblank(), but it's not guaranteed. */
9d116dd7 771#ifdef EBCDIC
b8c5462f 772# define isALNUMC(c) isalnum(c)
271b5cb3 773# define isALPHA(c) isalpha(c)
0bc70f7f 774# define isBLANK(c) ((c) == ' ' || (c) == '\t' || NATIVE_TO_UNI(c) == 0xA0)
b8c5462f 775# define isCNTRL(c) iscntrl(c)
9916188f 776# define isDIGIT(c) isdigit(c)
b8c5462f 777# define isGRAPH(c) isgraph(c)
8a58bdcf 778# define isIDFIRST(c) (isALPHA(c) || (c) == '_')
7d79683e 779# define isLOWER(c) islower(c)
9d116dd7 780# define isPRINT(c) isprint(c)
56e4b611 781# define isPSXSPC(c) isspace(c)
b8c5462f 782# define isPUNCT(c) ispunct(c)
174b5e76 783# define isSPACE(c) (isPSXSPC(c) && (c) != '\v')
7d79683e 784# define isUPPER(c) isupper(c)
b8c5462f 785# define isXDIGIT(c) isxdigit(c)
8a58bdcf 786# define isWORDCHAR(c) (isalnum(c) || (c) == '_')
9d116dd7 787# define toLOWER(c) tolower(c)
7d79683e 788# define toUPPER(c) toupper(c)
ec79cfbb 789#else /* Not EBCDIC: ASCII-only matching */
e58b1dde 790# define isALNUMC(c) isALNUMC_A(c) /* Mnemonic: "C's alnum" = alpha + digit */
271b5cb3 791# define isALPHA(c) isALPHA_A(c)
ec79cfbb
KW
792# define isBLANK(c) isBLANK_A(c)
793# define isCNTRL(c) isCNTRL_A(c)
794# define isDIGIT(c) isDIGIT_A(c)
795# define isGRAPH(c) isGRAPH_A(c)
8a58bdcf 796# define isIDFIRST(c) isIDFIRST_A(c)
ec79cfbb
KW
797# define isLOWER(c) isLOWER_A(c)
798# define isPRINT(c) isPRINT_A(c)
799# define isPSXSPC(c) isPSXSPC_A(c)
800# define isPUNCT(c) isPUNCT_A(c)
801# define isSPACE(c) isSPACE_A(c)
802# define isUPPER(c) isUPPER_A(c)
8a58bdcf 803# define isWORDCHAR(c) isWORDCHAR_A(c)
ec79cfbb 804# define isXDIGIT(c) isXDIGIT_A(c)
00f254e2 805
8a58bdcf
KW
806 /* ASCII casing. These could also be written as
807 #define toLOWER(c) (isASCII(c) ? toLOWER_LATIN1(c) : (c))
808 #define toUPPER(c) (isASCII(c) ? toUPPER_LATIN1_MOD(c) : (c))
809 which uses table lookup and mask instead of subtraction. (This would
810 work because the _MOD does not apply in the ASCII range) */
9d116dd7 811# define toLOWER(c) (isUPPER(c) ? (c) + ('a' - 'A') : (c))
7d79683e 812# define toUPPER(c) (isLOWER(c) ? (c) - ('a' - 'A') : (c))
9d116dd7 813#endif
bbce6d69 814
1a0901db
KW
815
816/* Use table lookup for speed; return error character for input
817 * out-of-range */
818#define toLOWER_LATIN1(c) (FITS_IN_8_BITS(c) \
819 ? UNI_TO_NATIVE(PL_latin1_lc[ \
820 NATIVE_TO_UNI( (U8) (c)) ]) \
821 : UNICODE_REPLACEMENT)
822/* Modified uc. Is correct uc except for three non-ascii chars which are
823 * all mapped to one of them, and these need special handling; error
824 * character for input out-of-range */
825#define toUPPER_LATIN1_MOD(c) (FITS_IN_8_BITS(c) \
826 ? UNI_TO_NATIVE(PL_mod_latin1_uc[ \
827 NATIVE_TO_UNI( (U8) (c)) ]) \
828 : UNICODE_REPLACEMENT)
829
bbce6d69
PP
830#ifdef USE_NEXT_CTYPE
831
832# define isALNUM_LC(c) \
37bd1396 833 (NXIsAlNum((unsigned int)(c)) || (char)(c) == '_')
ff68c719
PP
834# define isIDFIRST_LC(c) \
835 (NXIsAlpha((unsigned int)(c)) || (char)(c) == '_')
836# define isALPHA_LC(c) NXIsAlpha((unsigned int)(c))
07315176 837# define isASCII_LC(c) isASCII((unsigned int)(c))
67addccf 838# define isBLANK_LC(c) isBLANK((unsigned int)(c))
ff68c719
PP
839# define isSPACE_LC(c) NXIsSpace((unsigned int)(c))
840# define isDIGIT_LC(c) NXIsDigit((unsigned int)(c))
841# define isUPPER_LC(c) NXIsUpper((unsigned int)(c))
842# define isLOWER_LC(c) NXIsLower((unsigned int)(c))
37bd1396 843# define isALNUMC_LC(c) NXIsAlNum((unsigned int)(c))
b8c5462f
JH
844# define isCNTRL_LC(c) NXIsCntrl((unsigned int)(c))
845# define isGRAPH_LC(c) NXIsGraph((unsigned int)(c))
ff68c719 846# define isPRINT_LC(c) NXIsPrint((unsigned int)(c))
b8c5462f 847# define isPUNCT_LC(c) NXIsPunct((unsigned int)(c))
ff68c719
PP
848# define toUPPER_LC(c) NXToUpper((unsigned int)(c))
849# define toLOWER_LC(c) NXToLower((unsigned int)(c))
bbce6d69
PP
850
851#else /* !USE_NEXT_CTYPE */
b8c5462f 852
bbce6d69
PP
853# if defined(CTYPE256) || (!defined(isascii) && !defined(HAS_ISASCII))
854
851131d3
KW
855/* Note that the foo_LC() macros in this case generally are defined only on
856 * code points 0-256, and give undefined, unwarned results if called with
857 * values outside that range */
858
b8c5462f 859# define isALNUM_LC(c) (isalnum((unsigned char)(c)) || (char)(c) == '_')
ff68c719 860# define isIDFIRST_LC(c) (isalpha((unsigned char)(c)) || (char)(c) == '_')
bbce6d69 861# define isALPHA_LC(c) isalpha((unsigned char)(c))
07315176
KW
862# ifdef HAS_ISASCII
863# define isASCII_LC(c) isascii((unsigned char)(c))
864# else
865# define isASCII_LC(c) isASCII((unsigned char)(c))
866# endif
67addccf
KW
867# ifdef HAS_ISBLANK
868# define isBLANK_LC(c) isblank((unsigned char)(c))
869# else
870# define isBLANK_LC(c) isBLANK((unsigned char)(c))
871# endif
bbce6d69
PP
872# define isSPACE_LC(c) isspace((unsigned char)(c))
873# define isDIGIT_LC(c) isdigit((unsigned char)(c))
874# define isUPPER_LC(c) isupper((unsigned char)(c))
875# define isLOWER_LC(c) islower((unsigned char)(c))
b8c5462f
JH
876# define isALNUMC_LC(c) isalnum((unsigned char)(c))
877# define isCNTRL_LC(c) iscntrl((unsigned char)(c))
878# define isGRAPH_LC(c) isgraph((unsigned char)(c))
bbce6d69 879# define isPRINT_LC(c) isprint((unsigned char)(c))
b8c5462f 880# define isPUNCT_LC(c) ispunct((unsigned char)(c))
bbce6d69
PP
881# define toUPPER_LC(c) toupper((unsigned char)(c))
882# define toLOWER_LC(c) tolower((unsigned char)(c))
883
884# else
885
3609ea0d 886# define isALNUM_LC(c) (isascii(c) && (isalnum(c) || (c) == '_'))
bbce6d69
PP
887# define isIDFIRST_LC(c) (isascii(c) && (isalpha(c) || (c) == '_'))
888# define isALPHA_LC(c) (isascii(c) && isalpha(c))
07315176 889# define isASCII_LC(c) isascii(c)
67addccf
KW
890# ifdef HAS_ISBLANK
891# define isBLANK_LC(c) (isascii(c) && isblank(c))
892# else
893# define isBLANK_LC(c) isBLANK(c)
894# endif
bbce6d69
PP
895# define isSPACE_LC(c) (isascii(c) && isspace(c))
896# define isDIGIT_LC(c) (isascii(c) && isdigit(c))
897# define isUPPER_LC(c) (isascii(c) && isupper(c))
898# define isLOWER_LC(c) (isascii(c) && islower(c))
b8c5462f
JH
899# define isALNUMC_LC(c) (isascii(c) && isalnum(c))
900# define isCNTRL_LC(c) (isascii(c) && iscntrl(c))
901# define isGRAPH_LC(c) (isascii(c) && isgraph(c))
bbce6d69 902# define isPRINT_LC(c) (isascii(c) && isprint(c))
b8c5462f 903# define isPUNCT_LC(c) (isascii(c) && ispunct(c))
bbce6d69
PP
904# define toUPPER_LC(c) toupper(c)
905# define toLOWER_LC(c) tolower(c)
906
907# endif
a0d0e21e 908#endif /* USE_NEXT_CTYPE */
55204971 909
aaa51d5e 910#define isPSXSPC_LC(c) (isSPACE_LC(c) || (c) == '\v')
aaa51d5e 911
eb0052d1
KW
912/* For use in the macros just below. If the input is Latin1, use the Latin1
913 * (_L1) version of the macro; otherwise use the function. Won't compile if
914 * 'c' isn't unsigned, as won't match function prototype. The macros do bounds
915 * checking, so have duplicate checks here, so could create versions of the
916 * macros that don't, but experiments show that gcc optimizes them out anyway.
917 */
918#define generic_uni(macro, function, c) ((c) < 256 \
919 ? CAT2(macro, _L1)(c) \
920 : function(c))
181ab5fd 921/* Note that all ignore 'use bytes' */
eb0052d1
KW
922
923#define isALNUM_uni(c) generic_uni(isWORDCHAR, is_uni_alnum, c)
bdd8600f 924#define isBLANK_uni(c) generic_uni(isBLANK, is_uni_blank, c)
eb0052d1
KW
925#define isIDFIRST_uni(c) generic_uni(isIDFIRST, is_uni_idfirst, c)
926#define isALPHA_uni(c) generic_uni(isALPHA, is_uni_alpha, c)
927#define isSPACE_uni(c) generic_uni(isSPACE, is_uni_space, c)
928#define isDIGIT_uni(c) generic_uni(isDIGIT, is_uni_digit, c)
929#define isUPPER_uni(c) generic_uni(isUPPER, is_uni_upper, c)
930#define isLOWER_uni(c) generic_uni(isLOWER, is_uni_lower, c)
141d8bad 931#define isASCII_uni(c) isASCII(c)
7b952154 932/* All controls are in Latin1 */
8ad3d40f 933#define isCNTRL_uni(c) isCNTRL_L1(c)
eb0052d1
KW
934#define isGRAPH_uni(c) generic_uni(isGRAPH, is_uni_graph, c)
935#define isPRINT_uni(c) generic_uni(isPRINT, is_uni_print, c)
936#define isPUNCT_uni(c) generic_uni(isPUNCT, is_uni_punct, c)
937#define isXDIGIT_uni(c) generic_uni(isXDIGIT, is_uni_xdigit, c)
b18192e9
KW
938
939/* Posix and regular space differ only in U+000B, which is in Latin1 */
940#define isPSXSPC_uni(c) ((c) < 256 ? isPSXSPC_L1(c) : isSPACE_uni(c))
941
a2a2844f
JH
942#define toUPPER_uni(c,s,l) to_uni_upper(c,s,l)
943#define toTITLE_uni(c,s,l) to_uni_title(c,s,l)
944#define toLOWER_uni(c,s,l) to_uni_lower(c,s,l)
b8d68ded 945#define toFOLD_uni(c,s,l) to_uni_fold(c,s,l)
a0ed51b3 946
9041c2e3
NIS
947#define isALNUM_LC_uvchr(c) (c < 256 ? isALNUM_LC(c) : is_uni_alnum_lc(c))
948#define isIDFIRST_LC_uvchr(c) (c < 256 ? isIDFIRST_LC(c) : is_uni_idfirst_lc(c))
949#define isALPHA_LC_uvchr(c) (c < 256 ? isALPHA_LC(c) : is_uni_alpha_lc(c))
950#define isSPACE_LC_uvchr(c) (c < 256 ? isSPACE_LC(c) : is_uni_space_lc(c))
951#define isDIGIT_LC_uvchr(c) (c < 256 ? isDIGIT_LC(c) : is_uni_digit_lc(c))
952#define isUPPER_LC_uvchr(c) (c < 256 ? isUPPER_LC(c) : is_uni_upper_lc(c))
953#define isLOWER_LC_uvchr(c) (c < 256 ? isLOWER_LC(c) : is_uni_lower_lc(c))
9041c2e3
NIS
954#define isCNTRL_LC_uvchr(c) (c < 256 ? isCNTRL_LC(c) : is_uni_cntrl_lc(c))
955#define isGRAPH_LC_uvchr(c) (c < 256 ? isGRAPH_LC(c) : is_uni_graph_lc(c))
956#define isPRINT_LC_uvchr(c) (c < 256 ? isPRINT_LC(c) : is_uni_print_lc(c))
957#define isPUNCT_LC_uvchr(c) (c < 256 ? isPUNCT_LC(c) : is_uni_punct_lc(c))
a0ed51b3 958
aaa51d5e
JF
959#define isPSXSPC_LC_uni(c) (isSPACE_LC_uni(c) ||(c) == '\f')
960#define isBLANK_LC_uni(c) isBLANK(c) /* could be wrong */
961
c8362b00 962/* For use in the macros just below. If the input is ASCII, use the ASCII (_A)
7d7a6efc
KW
963 * version of the macro; if the input is in the upper Latin1 range, use the
964 * Latin1 (_L1) version of the macro, after converting from utf8; otherwise use
965 * the function. This relies on the fact that ASCII characters have the same
966 * representation whether utf8 or not */
967#define generic_utf8(macro, function, p) (isASCII(*(p)) \
b18192e9 968 ? CAT2(CAT2(macro,_),A)(*(p)) \
7d7a6efc
KW
969 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
970 ? CAT2(macro, _L1) \
971 (TWO_BYTE_UTF8_TO_UNI(*(p), \
972 *((p)+1))) \
973 : function(p))
c8362b00 974
4b88fb76
KW
975/* Note that all assume that the utf8 has been validated, and ignore 'use
976 * bytes' */
181ab5fd 977
c8362b00 978#define isALNUM_utf8(p) generic_utf8(isWORDCHAR, is_utf8_alnum, p)
c11ff943
KW
979/* To prevent S_scan_word in toke.c from hanging, we have to make sure that
980 * IDFIRST is an alnum. See
c8362b00 981 * http://rt.perl.org/rt3/Ticket/Display.html?id=74022 for more detail than you
9479a769
KW
982 * ever wanted to know about. XXX It is unclear if this should extend to
983 * isIDFIRST_uni() which it hasn't so far. (In the ASCII range, there isn't a
984 * difference.) This used to be not the XID version, but we decided to go with
985 * the more modern Unicode definition */
c8362b00
KW
986#define isIDFIRST_utf8(p) (isASCII(*(p)) \
987 ? isIDFIRST_A(*(p)) \
7d7a6efc
KW
988 : (UTF8_IS_DOWNGRADEABLE_START(*(p))) \
989 ? isIDFIRST_L1(TWO_BYTE_UTF8_TO_UNI(*(p), \
990 *((p)+1)))\
4244b029 991 : Perl__is_utf8__perl_idstart(aTHX_ p))
c8362b00
KW
992#define isIDCONT_utf8(p) generic_utf8(isWORDCHAR, is_utf8_xidcont, p)
993#define isALPHA_utf8(p) generic_utf8(isALPHA, is_utf8_alpha, p)
bdd8600f 994#define isBLANK_utf8(p) generic_utf8(isBLANK, is_utf8_blank, p)
c8362b00
KW
995#define isSPACE_utf8(p) generic_utf8(isSPACE, is_utf8_space, p)
996#define isDIGIT_utf8(p) generic_utf8(isDIGIT, is_utf8_digit, p)
997#define isUPPER_utf8(p) generic_utf8(isUPPER, is_utf8_upper, p)
998#define isLOWER_utf8(p) generic_utf8(isLOWER, is_utf8_lower, p)
141d8bad 999/* Because ASCII is invariant under utf8, the non-utf8 macro works */
bce3d7a4 1000#define isASCII_utf8(p) isASCII(*p)
c8362b00
KW
1001#define isCNTRL_utf8(p) generic_utf8(isCNTRL, is_utf8_cntrl, p)
1002#define isGRAPH_utf8(p) generic_utf8(isGRAPH, is_utf8_graph, p)
1003#define isPRINT_utf8(p) generic_utf8(isPRINT, is_utf8_print, p)
1004#define isPUNCT_utf8(p) generic_utf8(isPUNCT, is_utf8_punct, p)
1005#define isXDIGIT_utf8(p) generic_utf8(isXDIGIT, is_utf8_xdigit, p)
a2a2844f
JH
1006#define toUPPER_utf8(p,s,l) to_utf8_upper(p,s,l)
1007#define toTITLE_utf8(p,s,l) to_utf8_title(p,s,l)
1008#define toLOWER_utf8(p,s,l) to_utf8_lower(p,s,l)
a0ed51b3 1009
c8362b00
KW
1010/* Posix and regular space differ only in U+000B, which is in ASCII (and hence
1011 * Latin1 */
1012#define isPSXSPC_utf8(p) ((isASCII(*(p))) \
1013 ? isPSXSPC_A(*(p)) \
7d7a6efc
KW
1014 : (UTF8_IS_DOWNGRADEABLE_START(*(p)) \
1015 ? isPSXSPC_L1(TWO_BYTE_UTF8_TO_UNI(*(p), \
1016 *((p)+1)))\
1017 : isSPACE_utf8(p)))
4b88fb76
KW
1018#define isALNUM_LC_utf8(p) isALNUM_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1019#define isIDFIRST_LC_utf8(p) isIDFIRST_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1020#define isALPHA_LC_utf8(p) isALPHA_LC_uvchr(valid_utf8_to_uvchr(p, 0))
bdd8600f 1021#define isBLANK_LC_utf8(p) isBLANK_LC_uvchr(valid_utf8_to_uvchr(p, 0))
4b88fb76
KW
1022#define isSPACE_LC_utf8(p) isSPACE_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1023#define isDIGIT_LC_utf8(p) isDIGIT_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1024#define isUPPER_LC_utf8(p) isUPPER_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1025#define isLOWER_LC_utf8(p) isLOWER_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1026#define isALNUMC_LC_utf8(p) isALNUMC_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1027#define isCNTRL_LC_utf8(p) isCNTRL_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1028#define isGRAPH_LC_utf8(p) isGRAPH_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1029#define isPRINT_LC_utf8(p) isPRINT_LC_uvchr(valid_utf8_to_uvchr(p, 0))
1030#define isPUNCT_LC_utf8(p) isPUNCT_LC_uvchr(valid_utf8_to_uvchr(p, 0))
a0ed51b3 1031
aaa51d5e 1032#define isPSXSPC_LC_utf8(c) (isSPACE_LC_utf8(c) ||(c) == '\f')
aaa51d5e 1033
6ae9f32a
KW
1034/* This conversion works both ways, strangely enough. On EBCDIC platforms,
1035 * CTRL-@ is 0, CTRL-A is 1, etc, just like on ASCII */
1036# define toCTRL(c) (toUPPER(NATIVE_TO_UNI(c)) ^ 64)
bbce6d69 1037
dea28490
JJ
1038/* Line numbers are unsigned, 32 bits. */
1039typedef U32 line_t;
dea28490 1040#define NOLINE ((line_t) 4294967295UL)
378cc40b 1041
91152fc1
DG
1042/* Helpful alias for version prescan */
1043#define is_LAX_VERSION(a,b) \
1044 (a != Perl_prescan_version(aTHX_ a, FALSE, b, NULL, NULL, NULL, NULL))
1045
1046#define is_STRICT_VERSION(a,b) \
1047 (a != Perl_prescan_version(aTHX_ a, TRUE, b, NULL, NULL, NULL, NULL))
1048
1049#define BADVERSION(a,b,c) \
1050 if (b) { \
1051 *b = c; \
1052 } \
1053 return a;
8c52afec 1054
8e84507e 1055/*
ccfc67b7
JH
1056=head1 Memory Management
1057
a02a5408 1058=for apidoc Am|void|Newx|void* ptr|int nitems|type
954c1994
GS
1059The XSUB-writer's interface to the C C<malloc> function.
1060
c5008215
JC
1061In 5.9.3, Newx() and friends replace the older New() API, and drops
1062the first parameter, I<x>, a debug aid which allowed callers to identify
37b8b4c9 1063themselves. This aid has been superseded by a new build option,
d10b4965 1064PERL_MEM_LOG (see L<perlhacktips/PERL_MEM_LOG>). The older API is still
c5008215
JC
1065there for use in XS modules supporting older perls.
1066
a02a5408 1067=for apidoc Am|void|Newxc|void* ptr|int nitems|type|cast
954c1994 1068The XSUB-writer's interface to the C C<malloc> function, with
c5008215 1069cast. See also C<Newx>.
954c1994 1070
a02a5408 1071=for apidoc Am|void|Newxz|void* ptr|int nitems|type
954c1994 1072The XSUB-writer's interface to the C C<malloc> function. The allocated
c5008215 1073memory is zeroed with C<memzero>. See also C<Newx>.
a02a5408 1074
954c1994
GS
1075=for apidoc Am|void|Renew|void* ptr|int nitems|type
1076The XSUB-writer's interface to the C C<realloc> function.
1077
1078=for apidoc Am|void|Renewc|void* ptr|int nitems|type|cast
1079The XSUB-writer's interface to the C C<realloc> function, with
1080cast.
1081
49b8b560 1082=for apidoc Am|void|Safefree|void* ptr
954c1994
GS
1083The XSUB-writer's interface to the C C<free> function.
1084
1085=for apidoc Am|void|Move|void* src|void* dest|int nitems|type
1086The XSUB-writer's interface to the C C<memmove> function. The C<src> is the
1087source, C<dest> is the destination, C<nitems> is the number of items, and C<type> is
1088the type. Can do overlapping moves. See also C<Copy>.
1089
e90e2364
NC
1090=for apidoc Am|void *|MoveD|void* src|void* dest|int nitems|type
1091Like C<Move> but returns dest. Useful for encouraging compilers to tail-call
1092optimise.
1093
954c1994
GS
1094=for apidoc Am|void|Copy|void* src|void* dest|int nitems|type
1095The XSUB-writer's interface to the C C<memcpy> function. The C<src> is the
1096source, C<dest> is the destination, C<nitems> is the number of items, and C<type> is
1097the type. May fail on overlapping copies. See also C<Move>.
1098
e90e2364
NC
1099=for apidoc Am|void *|CopyD|void* src|void* dest|int nitems|type
1100
1101Like C<Copy> but returns dest. Useful for encouraging compilers to tail-call
1102optimise.
1103
954c1994
GS
1104=for apidoc Am|void|Zero|void* dest|int nitems|type
1105
1106The XSUB-writer's interface to the C C<memzero> function. The C<dest> is the
1107destination, C<nitems> is the number of items, and C<type> is the type.
1108
e90e2364
NC
1109=for apidoc Am|void *|ZeroD|void* dest|int nitems|type
1110
1111Like C<Zero> but returns dest. Useful for encouraging compilers to tail-call
1112optimise.
1113
da5d8dbb 1114=for apidoc Am|void|StructCopy|type *src|type *dest|type
4375e838 1115This is an architecture-independent macro to copy one structure to another.
954c1994 1116
7e337ee0
JH
1117=for apidoc Am|void|PoisonWith|void* dest|int nitems|type|U8 byte
1118
1119Fill up memory with a byte pattern (a byte repeated over and over
1120again) that hopefully catches attempts to access uninitialized memory.
1121
1122=for apidoc Am|void|PoisonNew|void* dest|int nitems|type
1123
1124PoisonWith(0xAB) for catching access to allocated but uninitialized memory.
1125
1c12ffb4 1126=for apidoc Am|void|PoisonFree|void* dest|int nitems|type
7e337ee0
JH
1127
1128PoisonWith(0xEF) for catching access to freed memory.
1129
9965345d
JH
1130=for apidoc Am|void|Poison|void* dest|int nitems|type
1131
7e337ee0 1132PoisonWith(0xEF) for catching access to freed memory.
9965345d
JH
1133
1134=cut */
954c1994 1135
561b68a9
SH
1136/* Maintained for backwards-compatibility only. Use newSV() instead. */
1137#ifndef PERL_CORE
ff06c60c 1138#define NEWSV(x,len) newSV(len)
561b68a9 1139#endif
ff06c60c 1140
19a94d75
JH
1141#define MEM_SIZE_MAX ((MEM_SIZE)~0)
1142
2720f793
JH
1143/* The +0.0 in MEM_WRAP_CHECK_ is an attempt to foil
1144 * overly eager compilers that will bleat about e.g.
1145 * (U16)n > (size_t)~0/sizeof(U16) always being false. */
27d5b266 1146#ifdef PERL_MALLOC_WRAP
fabdb6c0 1147#define MEM_WRAP_CHECK(n,t) MEM_WRAP_CHECK_1(n,t,PL_memory_wrap)
27d5b266 1148#define MEM_WRAP_CHECK_1(n,t,a) \
f1f66076 1149 (void)(sizeof(t) > 1 && ((MEM_SIZE)(n)+0.0) > MEM_SIZE_MAX/sizeof(t) && (Perl_croak_nocontext("%s",(a)),0))
8b44ba4c 1150#define MEM_WRAP_CHECK_(n,t) MEM_WRAP_CHECK(n,t),
27d5b266 1151
f1f66076 1152#define PERL_STRLEN_ROUNDUP(n) ((void)(((n) > MEM_SIZE_MAX - 2 * PERL_STRLEN_ROUNDUP_QUANTUM) ? (Perl_croak_nocontext("%s",PL_memory_wrap),0):0),((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
1936d2a7 1153
27d5b266
JH
1154#else
1155
410319be
NC
1156#define MEM_WRAP_CHECK(n,t)
1157#define MEM_WRAP_CHECK_1(n,t,a)
1158#define MEM_WRAP_CHECK_2(n,t,a,b)
8b44ba4c
NC
1159#define MEM_WRAP_CHECK_(n,t)
1160
6fff79ce 1161#define PERL_STRLEN_ROUNDUP(n) (((n-1+PERL_STRLEN_ROUNDUP_QUANTUM)&~((MEM_SIZE)PERL_STRLEN_ROUNDUP_QUANTUM-1)))
27d5b266 1162
1936d2a7 1163#endif
8b44ba4c 1164
fe4f188c 1165#ifdef PERL_MEM_LOG
46c6c7e2 1166/*
9f653bb5 1167 * If PERL_MEM_LOG is defined, all Newx()s, Renew()s, and Safefree()s
46c6c7e2
JH
1168 * go through functions, which are handy for debugging breakpoints, but
1169 * which more importantly get the immediate calling environment (file and
e352bcff
JH
1170 * line number, and C function name if available) passed in. This info can
1171 * then be used for logging the calls, for which one gets a sample
73d1d973 1172 * implementation unless -DPERL_MEM_LOG_NOIMPL is also defined.
3609ea0d 1173 *
46c6c7e2 1174 * Known problems:
94e892a6 1175 * - not all memory allocs get logged, only those
46c6c7e2 1176 * that go through Newx() and derivatives (while all
94e892a6 1177 * Safefrees do get logged)
46c6c7e2
JH
1178 * - __FILE__ and __LINE__ do not work everywhere
1179 * - __func__ or __FUNCTION__ even less so
1180 * - I think more goes on after the perlio frees but
1181 * the thing is that STDERR gets closed (as do all
1182 * the file descriptors)
1183 * - no deeper calling stack than the caller of the Newx()
1184 * or the kind, but do I look like a C reflection/introspection
1185 * utility to you?
1186 * - the function prototypes for the logging functions
1187 * probably should maybe be somewhere else than handy.h
1188 * - one could consider inlining (macrofying) the logging
1189 * for speed, but I am too lazy
1190 * - one could imagine recording the allocations in a hash,
1191 * (keyed by the allocation address?), and maintain that
1192 * through reallocs and frees, but how to do that without
1193 * any News() happening...?
73d1d973 1194 * - lots of -Ddefines to get useful/controllable output
b953482e 1195 * - lots of ENV reads
46c6c7e2
JH
1196 */
1197
12754f92 1198PERL_EXPORT_C Malloc_t Perl_mem_log_alloc(const UV n, const UV typesize, const char *type_name, Malloc_t newalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1199
12754f92 1200PERL_EXPORT_C Malloc_t Perl_mem_log_realloc(const UV n, const UV typesize, const char *type_name, Malloc_t oldalloc, Malloc_t newalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1201
12754f92 1202PERL_EXPORT_C Malloc_t Perl_mem_log_free(Malloc_t oldalloc, const char *filename, const int linenumber, const char *funcname);
46c6c7e2 1203
0b0ab801 1204# ifdef PERL_CORE
73d1d973 1205# ifndef PERL_MEM_LOG_NOIMPL
0b0ab801
MHM
1206enum mem_log_type {
1207 MLT_ALLOC,
1208 MLT_REALLOC,
d7a2c63c
MHM
1209 MLT_FREE,
1210 MLT_NEW_SV,
1211 MLT_DEL_SV
0b0ab801
MHM
1212};
1213# endif
12754f92 1214# if defined(PERL_IN_SV_C) /* those are only used in sv.c */
d7a2c63c
MHM
1215void Perl_mem_log_new_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
1216void Perl_mem_log_del_sv(const SV *sv, const char *filename, const int linenumber, const char *funcname);
12754f92 1217# endif
0b0ab801
MHM
1218# endif
1219
fe4f188c
JH
1220#endif
1221
1222#ifdef PERL_MEM_LOG
d1401ee9
MHM
1223#define MEM_LOG_ALLOC(n,t,a) Perl_mem_log_alloc(n,sizeof(t),STRINGIFY(t),a,__FILE__,__LINE__,FUNCTION__)
1224#define MEM_LOG_REALLOC(n,t,v,a) Perl_mem_log_realloc(n,sizeof(t),STRINGIFY(t),v,a,__FILE__,__LINE__,FUNCTION__)
46c6c7e2 1225#define MEM_LOG_FREE(a) Perl_mem_log_free(a,__FILE__,__LINE__,FUNCTION__)
fe4f188c
JH
1226#endif
1227
1228#ifndef MEM_LOG_ALLOC
1229#define MEM_LOG_ALLOC(n,t,a) (a)
1230#endif
1231#ifndef MEM_LOG_REALLOC
1232#define MEM_LOG_REALLOC(n,t,v,a) (a)
1233#endif
1234#ifndef MEM_LOG_FREE
1235#define MEM_LOG_FREE(a) (a)
1236#endif
1237
d1401ee9
MHM
1238#define Newx(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1239#define Newxc(v,n,t,c) (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_ALLOC(n,t,safemalloc((MEM_SIZE)((n)*sizeof(t))))))
1240#define Newxz(v,n,t) (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_ALLOC(n,t,safecalloc((n),sizeof(t)))))
a6f6820f
NC
1241
1242#ifndef PERL_CORE
a02a5408
JC
1243/* pre 5.9.x compatibility */
1244#define New(x,v,n,t) Newx(v,n,t)
1245#define Newc(x,v,n,t,c) Newxc(v,n,t,c)
4541904d 1246#define Newz(x,v,n,t) Newxz(v,n,t)
a6f6820f 1247#endif
a02a5408 1248
ff68c719 1249#define Renew(v,n,t) \
d1401ee9 1250 (v = (MEM_WRAP_CHECK_(n,t) (t*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
ff68c719 1251#define Renewc(v,n,t,c) \
d1401ee9 1252 (v = (MEM_WRAP_CHECK_(n,t) (c*)MEM_LOG_REALLOC(n,t,v,saferealloc((Malloc_t)(v),(MEM_SIZE)((n)*sizeof(t))))))
94010e71
NC
1253
1254#ifdef PERL_POISON
1255#define Safefree(d) \
06c0cc96 1256 ((d) ? (void)(safefree(MEM_LOG_FREE((Malloc_t)(d))), Poison(&(d), 1, Malloc_t)) : (void) 0)
94010e71 1257#else
fe4f188c 1258#define Safefree(d) safefree(MEM_LOG_FREE((Malloc_t)(d)))
94010e71 1259#endif
55497cff 1260
8b44ba4c
NC
1261#define Move(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1262#define Copy(s,d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
1263#define Zero(d,n,t) (MEM_WRAP_CHECK_(n,t) (void)memzero((char*)(d), (n) * sizeof(t)))
55497cff 1264
8b44ba4c
NC
1265#define MoveD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memmove((char*)(d),(const char*)(s), (n) * sizeof(t)))
1266#define CopyD(s,d,n,t) (MEM_WRAP_CHECK_(n,t) memcpy((char*)(d),(const char*)(s), (n) * sizeof(t)))
e90e2364 1267#ifdef HAS_MEMSET
8b44ba4c 1268#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)))
e90e2364 1269#else
8b44ba4c
NC
1270/* Using bzero(), which returns void. */
1271#define ZeroD(d,n,t) (MEM_WRAP_CHECK_(n,t) memzero((char*)(d), (n) * sizeof(t)),d)
e90e2364
NC
1272#endif
1273
7e337ee0
JH
1274#define PoisonWith(d,n,t,b) (MEM_WRAP_CHECK_(n,t) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)))
1275#define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB)
1276#define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF)
1277#define Poison(d,n,t) PoisonFree(d,n,t)
27d5b266 1278
2304df62 1279#ifdef USE_STRUCT_COPY
ff68c719 1280#define StructCopy(s,d,t) (*((t*)(d)) = *((t*)(s)))
bee1dbe2
LW
1281#else
1282#define StructCopy(s,d,t) Copy(s,d,1,t)
1283#endif
2cc61e15 1284
622913ab 1285#define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0]))
cd431fde 1286#define C_ARRAY_END(a) (a) + (sizeof(a)/sizeof((a)[0]))
622913ab 1287
2cc61e15
DD
1288#ifdef NEED_VA_COPY
1289# ifdef va_copy
1290# define Perl_va_copy(s, d) va_copy(d, s)
2cc61e15 1291# else
a1866d1b
JH
1292# if defined(__va_copy)
1293# define Perl_va_copy(s, d) __va_copy(d, s)
1294# else
1295# define Perl_va_copy(s, d) Copy(s, d, 1, va_list)
1296# endif
2cc61e15
DD
1297# endif
1298#endif
1299
472d47bc
SB
1300/* convenience debug macros */
1301#ifdef USE_ITHREADS
1302#define pTHX_FORMAT "Perl interpreter: 0x%p"
1303#define pTHX__FORMAT ", Perl interpreter: 0x%p"
f54cb97a
AL
1304#define pTHX_VALUE_ (void *)my_perl,
1305#define pTHX_VALUE (void *)my_perl
1306#define pTHX__VALUE_ ,(void *)my_perl,
1307#define pTHX__VALUE ,(void *)my_perl
472d47bc 1308#else
3609ea0d 1309#define pTHX_FORMAT
472d47bc 1310#define pTHX__FORMAT
3609ea0d 1311#define pTHX_VALUE_
472d47bc 1312#define pTHX_VALUE
3609ea0d 1313#define pTHX__VALUE_
472d47bc
SB
1314#define pTHX__VALUE
1315#endif /* USE_ITHREADS */
3609ea0d 1316
2acdbac1
NC
1317/* Perl_deprecate was not part of the public API, and did not have a deprecate()
1318 shortcut macro defined without -DPERL_CORE. Neither codesearch.google.com nor
1319 CPAN::Unpack show any users outside the core. */
1320#ifdef PERL_CORE
d1d15184 1321# define deprecate(s) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), "Use of " s " is deprecated")
2acdbac1
NC
1322#endif
1323
e9a8c099
MHM
1324/*
1325 * Local variables:
1326 * c-indentation-style: bsd
1327 * c-basic-offset: 4
14d04a33 1328 * indent-tabs-mode: nil
e9a8c099
MHM
1329 * End:
1330 *
14d04a33 1331 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 1332 */